Open Responses: What you need to know

Open Responses is a new, open inference standard designed to replace older formats like Chat Completion for modern agentic AI workloads. It extends OpenAI's Responses API to create a shared, interoperable format for developers, model providers, and routers. The standard is stateless by default, models streaming as semantic events, and formalizes the agentic loop of reasoning, tool invocation, and response generation. This allows models to autonomously handle multi-step tasks using both internal and external tools within a single request. The goal is to create a more consistent, capable, and standardized inference experience across the AI ecosystem.